Text version of the page

Parallel processing with ProCAST

on HP ProLiant platforms

CASTING

ESI Group / HP Alliance Increasing Casting Industry Productivity

HP and ESI Group are committed to developing and supplying a wide rangeof Linux-based products and servicestargeted specifically at the clusterenvironment for the casting industry.This shared commitment, coupled withHP's experience and leadership in theHPC market and Linux, is reflected inHP’s Unified Cluster Portfolio and the newparallel HP-MPI ProCAST solution.This trend towards increasingly powerfuland affordable hardware, combined withsophisticated and easier-to-use software,offers the casting industry : High computing performance,Hardware and software flexibility,High return on investment,Reliability through HP and ESI Groupcertification. Cutting costs and reducing time to market are two of the most pressing issues in the foundry industry today. Casting process simulation helps achieve these goals and is now widely used throughout the industry for process design, improved yield and better casting quality. While simplified solutions can be usedto quickly evaluate mold filling and solidification, the accurate prediction of complex flows and metallurgical phenomena, as provided by ProCAST, stillrequires significant computational times. Based on powerful Finite Element solvers and advanced specific optionsdeveloped with leading research institutes and industries, ProCAST provides a complete and accurate solution which covers a wide range of casting processes and alloys. With parallel processing on HP cluster platforms, foundries can significantly improve productivity, bringing products to market faster.Compute clusters are fast replacing the legacy shared memory parallel systems thathave been the traditional technology used in many high performance applications. Clusters can lower total cost of ownership and time to market by performing simulations even faster than giant supercomputers. The Unified Cluster Portfolio,with HP Cluster Platforms, allows you to quickly deploy clusters with ProLiantservers and popular interconnects, as well as provide support for Red Hat and SUSE Linux distributions and cluster management options such as XC System Software or Scali Connect.The ProCAST DMP (Distributed Memory Parallel) solution uses the latest availabletechnology, including dynamic domain decomposition and message passing communications, to provide the best performance on Linux clusters. To satisfy the requirements for an efficient, portable and flexible application, ProCAST uses the new HP-MPI technology to support parallel message passing on HP Linux clusters.

Running ProCAST on HP Linux clusters

The ProCAST DMP solution based on HP-MPI provides optimal scalability on HPProLiant Linux clusters running AMD Opteron or Intel
® Xeon™ EM64T.This solution offers the additional advantage of being independent from theinterconnect. Together with the flexibility of HP’s modular Unified Cluster Portfolio, you can design and deploy the high performance Linux solution that meet your needs: Select your proliant architecture: AMD Opteron, or Intel
® Xeon™ EM64T,Select node and server upgrades,Select your interconnect: Quadrics, Myricom, InfiniBand, Gigabit Ethernet, Configure up to 512 nodes.
HP ProLiant Linux cluster

www.esi-group.com